Part 3. Imatinib for Chronic Myeloid Leukemia (CML) (continued)

Discussion

In this section we summarize the findings of the review in terms of answering the key questions initially posed, and then discuss the clinical and research implications of these data.

CML is a rare hematological cancer that affects <5,000 Americans yearly. An excessive number of abnormal white blood cells are produced that eventually take over the body's ability to produce normal cells. In at least 95 percent of cases, CML starts with the formation of the Philadelphia chromosome (Ph), also known as the 9;22 translocation that forms the BCR-ABL gene. BCR-ABL is transcribed into mRNA and then translated into the BCR-ABL tyrosine kinase protein. This tyrosine kinase is a continuously active protein that sends the cancer signal of uncontrolled cell division. Imatinib binds to the BCR-ABL tyrosine kinase protein and turns off this signal.

There are three clinical phases of CML-chronic phase, accelerated phase, and blastic phase/blast crisis. These phases are characterized by their tumor aggressiveness and prognosis. Therapeutic options include imatinib, interferon alpha with or without cytarabine, hydroxyurea, busulfan, other conventional chemotherapies, and stem cell transplantation (bone marrow transplantation, SCT). Allogeneic SCT is the only curative treatment for CML, however it is only available for 20-25 percent of patients due to lack of a suitable donor;¹⁴⁷ 15-30 percent treatment-related mortality can be expected with SCT.¹⁷

Key Questions

1. In patients with chronic myeloid leukemia, what is the effect of imatinib compared to interferon alpha or best supportive care on overall survival, disease free survival, remission rates (PR, CHR, cytogenetic remission), and quality of life (QOL)?

There is convincing evidence of the efficacy of imatinib for CML in all clinical settings as described in the matrix below. For many of these studies the results are still early and median survival has not been reached. This is especially true for those studies of CP CML. Thus, Complete CR (CCR) rates are compared across studies, as Complete CR is a major indictor of tumor response, is correlated with PFS and OS as demonstrated in Table 12, and is a major goal of therapy.²²

The most compelling evidence for the efficacy of imatinib is the IRIS trial, an international multi-center phase III trial of imatinib vs. interferon plus cytarabine as initial therapy for newly diagnosed chronic phase CML.⁵⁹ A previous phase III study by Guilhot, et al. had demonstrated that interferon plus cytarabine rendered superior cytogenetic response and survival when compared to interferon alone.⁴² Another phase III study by Baccarini, et al. of interferon vs. interferon plus cytarabine was more equivocal with interferon plus cytarabine yielding better cytogenetic responses but similar survival.¹⁴⁹ Complete CR rates were slightly better in the Guilhot study than the Baccarini study (15 percent vs. 8 percent, respectively). Thus, the IRIS comparison group of interferon plus cytarabine is as good as interferon alone, if not better. Use of the interferon plus cytarabine arm from the Guilhot study as a baseline comparator when needed is also reasonable.

In the IRIS study, imatinib was clearly superior to interferon plus cytarabine in terms of cytogenetic response (74 percent vs. 9 percent),⁵⁹ molecular response (42% vs. 13% of those with Complete CR at 6 months),¹¹ PFS (92% vs. 74% at 18 months),⁵⁹ and QOL (TOI 84.4 vs. 67.7).^60,61 Estimates of OS were not significantly different between imatinib and interferon plus cytarabine in the original IRIS publication.⁵⁹ Since 58 percent of participants on the interferon plus cytarabine arm crossed over to imatinib in this trial, estimates of OS for the individual groups were difficult. In a followup report on the IRIS trial, the 30-month OS for imatinib was 95 percent.⁶² This compares favorably to the previously reported 36-month OS rates for interferon plus cytarabine of 86 percent in the Guilhot study.⁴² QOL was studied as part of the IRIS trial, and patients receiving imatinib had significantly better total QOL, social/family well-being, and emotional well-being (Table 9).^60,61 Pasquini el al. reported similar findings in a Phase II trial conducted in Brazil.⁷⁶

There were some criticisms of the IRIS trial. Most notably, the overall mean dose intensity on the interferon plus cytarabine arm was only 58 percent of the target dose, with the dose intensity of the imatinib arm 97 percent of target.¹⁴⁷ This compares similarly to the Guilhot, et al. trial of interferon vs. interferon plus cytarabine where only 57 percent achieved the target dose intensity with interferon.⁴² The Baccarini study reported higher rates of achieving target dose intensity with interferon (70 percent),¹⁴⁹ but did not report different survival rates than those seen with the Guilhot, et al. trial.¹⁴⁷ The other main criticism of the IRIS trial is that PFS was calculated using loss of CHR, loss of Major CR, or increases in WBC as criteria for progression.¹⁴⁷ This criticism is reflective of the variability in definition of disease progression in CML. For this reason, comparison of more uniform endpoints across trials such as Complete CR or OS may be a more objective measure of relative efficacy.

Efficacy is clearly different by phase of disease and timing within the treatment algorithm, as reflected in Figure 6. Earlier phases and patients treated in the first-line setting had the highest response rates. CP patients treated earlier in the course (i.e., <1 year from diagnosis) had better response rates with imatinib than those treated later in the CP period.⁴⁴ In the post-interferon setting, the reason that the interferon was discontinued influenced response rates.² Regardless, significant Complete CR rates are seen with imatinib in all treatment settings, including patients who are heavily pre-treated with myelotoxic chemotherapy with or without SCT. The response rates for the heavily pre-treated CP patients are similar to those of the interferon-refractory or intolerant CP patients. The historic control group for the interferon-refractory or intolerant CP patients likely reflects the same or better response rates than would an appropriate control group for the heavily pre-treated CP patients; this group has been used for the comparator group in the heavily-pretreated CP setting.

The AP and BP studies do not report comparator groups, however previous studies suggest that fewer than 5 percent of AP patients achieve a Major CR with interferon.¹⁵⁰ The Complete CR rate for AP treated with interferon can therefore be expected to be lower than 5 percent, and BP lower yet. Studies identified in this review reported Complete CR rates with imatinib of 11-19 percent for AP and 0-10 percent for BP (Figure 6). One year survival rates of 74 percent (95 percent CI 68-81 percent) for AP patients treated with imatinib compare favorably to the historic 6-18 month median life expectancy described in Figure 2.⁸⁷ Similarly, the median OS of 6.5-7 months for BP patients treated with imatinib is longer than the historic prognosis of 3-6 months.²

An important limitation to the assessment of efficacy is that many of the studies cited have overlapping populations. This does not necessarily subtract from the value of the analysis as the different reports and studies are usually addressing different issues, but needs to be kept in mind when considering sample sizes quotes. The estimation of efficacy and predictors of response is also limited by the rapidly evolving nature of this field—making it difficult to ensure that an evidence report is up-to-date after an arbitrary evidence review date.

Other important issues of imatinib efficacy include timing of effect, appropriate dose, and relationship to SCT. Efficacy analyses should be considered in terms of duration of exposure to imatinib. In the setting of newly diagnosed CP CML, molecular response rates to imatinib increased steadily over the first two years on imatinib and then did not change substantially after 24 months.³² Complete CR rates on imatinib increased for at least 12 months after initiation of the drug,^8,63,70,73 whereas Complete CRs did not increase after 6 months on interferon-based therapies. Some authors have argued that Complete CR rates do not increase after 6 months on imatinib,²² however this current review demonstrated that they continue to increase for up to 12 months and that periods after 12 months have been poorly studied. Nonetheless, achieving molecular and cytogenetic responses were beneficial no matter how long it took to get there (Tables 12 and 13).

Patients who achieved an early response as minimally defined by either molecular response by 4 weeks or some cytogenetic response by 3 months had better PFS and OS.^48,72 The exact milestone cut-off that should be followed is unclear. Cytogenetic response milestones have been investigated at 3 and 6 months predominantly (Table 12), although changes can be identified out to 12 months (Table 3). Molecular response milestones have been investigated for 4 weeks, 3 months, 6 months, and 12 months (Table 13). These milestones can be used to identify patients who have had a suboptimal response to imatinib. Failure to achieve a significant cytogenetic response (Major or Complete) by 6-12 months is one criteria for suboptimal response that may indicate an increased dose of imatinib or shift in treatment plan. Similarly, molecular milestones are starting to be used when such laboratory facilities are available. Failure to achieve a >2 log reduction in the number of BCR-ABL mRNA transcripts by 3-6 months could be considered evidence of suboptimal response;³³ failure to achieve a >3 log reduction by 6-12 months could be considered suboptimal.¹¹ These analyses were primarily conducted with patients receiving 400 mg imatinib daily.

Starting imatinib doses are usually 400 mg daily for CP and 600 mg daily for AP and BP.¹⁴⁸ In accordance with FDA recommendations based upon the IRIS study, imatinib is administered daily at a dose of 400 mg in newly diagnosed CP patients.¹⁴⁷ Patients not achieving a CHR at 3 months or a Major CR at 12 months may be escalated to 400 mg twice daily. For Grade 2 non-hematologic toxicity, imatinib is withheld until toxicity resolves. After resolution of grade 2 toxicity, the drug is resumed at 400 mg daily. After resolution of grade 3 or 4 toxicity, the drug is resumed at 300 mg daily. There is clearly a dose response relationship with imatinib.¹⁴⁸ Several studies in different clinical settings support the additional therapeutic advantage of increasing to 800 mg a day. Imatinib resistance or CML relapses at 400 mg can be overcome by increasing to 800 mg, as described in the CP-interferon refractory setting.⁷¹ Further, starting at the 600 or 800 mg dose may induce more Complete CRs,^65,69 but with more adverse events (Table 10).

The IRIS trial also demonstrated that imatinib was tolerable and efficacious after progression on interferon-based therapy, and that patients receiving imatinib could still go on to SCT.⁵⁹ Numerous studies presented in Tables 3, 4, 7, and 8 also support these findings. Imatinib is effective and well tolerated in the setting of disease relapse after SCT,^119-121 and it does not preclude a patient from receiving a SCT.

Given the genetic variability of the American population, an important question for targeted drugs such as imatinib is whether the clinical research findings are limited to a specific portion of the population. Ethic and racial studies related to imatinib efficacy are few. IRIS was an international multi-site trial involving patients from at least 15 countries in North America, Europe, Australia and New Zealand.⁵⁹ Deshmukh and colleagues presented a similar imatinib experience to that reported in other studies evaluating a population recruited exclusively in India.⁸⁶ The Pasquini, et al. study involved participants recruited exclusively in Brazil and reported similar QOL findings to that described in IRIS.⁷⁶ Meanwhile, a retrospective chart review by George and colleagues of 26 patients from the Chicago area suggested that non-Caucasian patients had poorer response rates to imatinib than Caucasians (Table 1d).¹²³ Complete CR was achieved in 100% of Caucasians (6/6) and 14% of non-Caucasians (2/14). Considering all of these studies, it appears that imatinib has efficacy across genetically diverse populations, however given the findings of George, et al., further studies are needed, especially in the United States.

Is there a differential effect of imatinib for patients who are >65 years of age? Two abstracts were presented at the 2004 American Society of Hematology meeting that addressed this question (Table 1d).^124,125 Both studies suggested that imatinib was efficacious and well tolerated in patients >65 or 70 years of age, although less so than younger patients. The study by Bassi, et al. suggested that patients >65 years had significantly more adverse events than those <65 and therefore poorer tolerance of imatinib and fewer Complete CRs (36% vs. 57%, p=0.001).

Does imatinib lead to "cure?" Defining "cure" in CML is difficult. Even when imatinib-treated patients are in Complete CR, evidence of CML can be found.^2,7,10,35 Blast crisis can still occur in patients who developed a Complete CR on imatinib.¹²² Complete remission with imatinib in CML may be a conversion to a low grade chronic disease with continuous potential for relapse over the long term. For this reason, the debate between imatinib vs. SCT in early chronic phase when possible continues.

Finally, this review of efficacy is based upon a systematic review of prospective studies that met the criteria for inclusion. Efficacy summaries reflect an overview of statistically significant reported findings, and neither reflect review of other literature nor current clinical practice. The field is evolving so quickly that such a review quickly becomes outdated and regular updates are important.

2. In patients with chronic myeloid leukemia, what is the effect of imatinib compared to interferon alpha or best supportive care on adverse effects, tolerability, and compliance with treatment?

Imatinib has far fewer adverse effects (any grade and grade 3/4) compared with interferon. In the IRIS trial, imatinib most commonly caused neutropenia (61 percent), thrombocytopenia (57 percent), superficial edema (56 percent), nausea (44 percent), and abnormal liver function results (43 percent).59 Interferon plus cytarabine most commonly caused thrombocytopenia (79 percent), abnormal liver function results (74 percent), neutropenia (67 percent), fatigue (66 percent), nausea (61 percent), anemia (55 percent), and headache (43 percent). The incidence of grade 3/4 side effects was primarily hematological with imatinib (neutropenia 14 percent and thrombocytopenia 8 percent) whereas interferon plus cytarabine included fatigue (24 percent) and hematological (neutropenia 25 percent and thrombocytopenia 17 percent). The incidence of side effects increased with imatinib dose and phase of illness, as expected (Table 10). In particular, the hematologic side effects increased with advancing phases of illness. As demonstrated by Sneed, et al., Grade 3/4 myelosuppression predicts poorer tumor responses with imatinib, especially when the myelosuppression lasts for longer than 2 weeks (Table 15).¹¹⁰

Compliance with imatinib was not formally presented in the studies reviewed. Discussions with authors revealed that there is a forthcoming report investigating adherence to imatinib therapy using prescription data for a total of 4043 imatinib-treated patients tracked over 14 months¹⁵¹. Overall, the compliance rate was approximately 75 percent, and persistent continuation on therapy averaged 256 days of therapy over 12 months. Suboptimal adherence to imatinib therapy may be an under-recognized problem that requires active monitoring by healthcare professionals.

3. What patient or tumor characteristics distinguish treatment responders from non-responders and have potential to be used to target therapy? In addressing this question, we will focus on the following: (1) predictive patient or tumor characteristics that are related to the mechanism of action of the drug (i.e., molecular target; performance status, while a powerful predictor of outcome, is not related to mechanism of action); (2) candidates for diagnostic testing (even if not commercially or clinically available currently (e.g., PCR)); and, (3) patient or tumor characteristics that are associated with clinically important differences in treatment response.

As presented in the Introduction (Chapter 1), there is clear correlation between clinical prognostic factors (e.g., phase of disease, previous treatment, Sokal score, splenomegaly, percentage of blasts in the peripheral blood) and tumor response or survival with imatinib. These known prognostic factors can be used to identify high risk and low risk patients in the setting of imatinib therapy in a similar manner to other treatment settings. A full review of the hazard ratios for these clinical prognostic factors was outside the scope of this review. Here we concentrate on molecular factors that predict response to imatinib and are likely to be related to the targeted action of the drug.

Prognostic factors were divided into 5 groups:

1A) DNA factors assessed at the start of therapy.
1B) DNA factors monitored during therapy.
2) Production of the RNA message.
3) Interaction between the tyrosine kinase protein and imatinib.
4) Other factors (Tables 11-15).

Many of these have already been reviewed in the preceding section on efficacy. Additional observations are presented here.

At the start of therapy, patients with a high burden of disease as evidenced by 90-100 percent of Ph+ metaphases during cytogenetic analysis or more CD34+ cells in the bone marrow were more likely to have a poor tumor response and decreased overall survival. Similarly, evidence of clonal evolution (complex cytogenetics) in the accelerated or blastic phases of illness predicted poorer survival and increased risk of tumor progression. Cytogenetic clonal evolution was a significant predictor of risk of relapse and shortened survival, but did not consistently predict disease response. Evidence of chromosome 9 deletions predicted poorer PFS but not OS. Once imatinib therapy was started, both evidence of cytogenetic response and reduction in the numbers of CD34+ cells in the bone marrow predicted improved PFS and OS. Cytogenetic response can be used as a surrogate marker of overall CML tumor response.

Factors that relate to production of the RNA message and that predict tumor response were highlighted in the efficacy discussion. BCR-ABL mRNA transcript levels measured before therapy starts are not predictive of outcome. Molecular response using Q-RT-PCR predicts survival and durability of the tumor response; it can be used as a surrogate marker of tumor response. When the log reduction was >2 at 3 or 6 months, patients had better PFS; similarly, when the log reduction was >3 at 12 months, patients had better PFS. Reduction in the BCR-ABL/ABL ratio to <50 percent at 4 weeks was also predictive of better PFS. Recently authors have suggested the need to rationally test different algorithms using molecular monitoring at defined timepoints.¹³²

All of the predictors just described are currently available for clinical use. In particular, cytogenetic analysis including analysis of chromosome 9 is widely available. A recent abstract indicates that peripheral blood FISH analysis is possible, but it is inferior to bone marrow samples or RT-PCR.¹²⁷ Analysis of CD34+ cells by flow cytometry is available through most reference laboratories. Reliable Q-RT-PCR for molecular monitoring is available through specialized facilities and centralized laboratories, and may not be an option for all patients at present.²²

Newer analyses looking at genetic profiles using microarrays are in development. McLean and colleagues demonstrated that they could identify a microarray pattern characteristic of tumor response in CP CML.¹⁵² While not currently ready for widespread use as a diagnostic test, such genetic profiling has the future potential to assist in the identification of individuals likely to respond or not respond to imatinib.¹²⁹ Similarly, individual genes associated with drug resistance have been identified; overexpression of MRP-1 was correlated with tumor response to imatinib.¹⁰² These studies are preliminary and not ready for clinical application, but do suggest that genetic profiles or RT-PCR analyses of the expression of individual genes other than BCR-ABL may be used in the future to assist in tailoring the use of imatinib for individual patients.

There is an evolving literature on the molecular mechanisms of imatinib resistance at the protein level.^146,153 The majority of this literature did not meet the criteria for this review because quantitative correlations with clinical outcomes were not presented. Mutations in the tyrosine kinase domain of BCR-ABL that lead to changes in the protein may disturb imatinib binding and therefore lead to poorer tumor response with imatinib. Changes of particular interest are those that lead to protein alterations in the p-loop where ATP binds and the protein pocket where imatinib binds. Such mutations that affect imatinib binding may make the drug less efficacious. Thus far the evidence for direct clinical impact has been scant. Shah and colleagues demonstrated how more CML patients with mutations in the binding domain progressed than those without mutations.¹⁰⁸ In four abstracts presented at the American Society of Hematology meeting in December 2004, it was suggested that ABL, p-loop and binding pocket mutations were predictive of disease progression or aggressiveness, [Soverini, 2004 #858; Corm, 2004 #846; Deininger, 2004 #844; Hochhaus, 2004 #168] while a fifth abstract suggested that these did not correlate with outcome.¹¹² Ideally, patients who are unlikely to have a good response to imatinib due to such mutations would be identified early and transitioned to more appropriate therapy. Some groups have used molecular monitoring to predict mutational status.¹⁵⁴ Analysis of 214 IRIS participants treated with imatinib revealed that 61 percent of the 56 patients with a >2-fold increase in BCR-ABL mRNA transcript levels had mutations while only 0.6 percent of the 158 with stable transcript levels had mutations.

This work on mutations that lead to altered imatinib binding and efficacy is still in development, both in terms of identification of the important mutations and their clinical impact. In order for it to have widespread clinical applicability there must be practical methods of detecting protein mutations. Soverini, et al. recently described a denaturing High Performance Liquid Chromotography (HPLC) method to screen for ABL point mutations that may make routine detection of mutations more practical.¹²⁶

Finally, White and colleagues have described an in vitro assay to predict imatinib's ability to inhibit phosphorylation of the adaptor protein Crkl.¹⁴⁵ This assay could be used to predict those CML likely to achieve a Major MR at 12 months before imatinib was started. This work is in an early phase and has not been widely tested, but provides another opportunity to identify patients likely to respond to imatinib and those who may need to transition to other therapies.

This work can be summarized as in Figure 7.

This is a rapidly evolving area and new data are constantly emerging. This current review only reflects the landscape to June 2005. Some of these data will become more or less useful as new information is uncovered. New predictors are likely to be defined.

Current State of Clinical Use

According to the National Comprehensive Cancer Network (NCCN) guideline dated November 23, 2004, imatinib is the standard of care as first-line therapy for CP CML when patients are not eligible for SCT.⁵⁶ This recommendation of imatinib as first-line therapy is stronger than the previous NCCN guideline which presented imatinib and interferon-based therapy as more equal options. When patients are eligible for SCT, the choice of first-line therapy with imatinib or transplant is still under debate.

The recommended starting dose is 400 mg. The NCCN guideline recommends that therapy is modified if a CHR is not obtained by 3 months. Modification options include reconsideration of SCT, clinical trials, increasing the imatinib to 600-800 mg, or interferon with or without cytarabine. For patients who obtained a CHR at 3 months, 6 month evaluation should include cytogenetic analysis. Patients who achieve at least a Minor CR at 6 months should continue at their current dose or increase to 600-800 mg as tolerated. Potential therapy modifications for patients who do not achieve at least a Minor CR by 6 months again include reconsideration of SCT, clinical trials, increasing the imatinib to 600-800 mg, or interferon with or without cytarabine. For patients who achieve at least a Minor CR at 6 months, 12 month evaluation should again include cytogenetic analysis. Those in Complete CR should continue imatinib at the current dose. Those in Major CR should be increased to 600-800 mg as tolerated, and those in Minor or no CR should proceed with therapy modification or continue imatinib with the goal of maintaining hematologic remission only. The option to start patients out at higher doses of imatinib is presented.

The NCCN guideline recommends bone marrow cytogenetic analysis even if FISH or Q-RT-PCR are available, because cytogenetic findings including clonal evolution may indicate the need to consider other treatment strategies (e.g., clinical trial, increased imatinib dose). Management strategies in the setting of chromosome 9 deletions are not discussed nor is the role of molecular monitoring.

According to the National Cancer Institute (NCI) clinical guide at www.cancer.gov, the timing and role of imatinib for newly diagnosed CP CML are not as clear.¹⁹ This review was most recently updated in February 2005. Particular questions raised by the NCI reviewers include the following:

• What is the best dose of imatinib and should it be combined with other agents (such as interferon alfa and/or cytarabine)?
• What is the role of allogeneic stem cell transplantation for younger, eligible patients, and should it be offered before or after initiation of imatinib?
• Will transplantation be more or equally efficacious before or after failure on imatinib?
• Will responses on imatinib be durable for many years, or will responses be short-lived and the relapsing disease be more difficult to control?

Both the NCCN and NCI guidelines are less clear about the optimal management of newly diagnosed AP or BP. Patients with newly diagnosed AP may be enrolled in a clinical trial, undergo SCT, be treated with imatinib, or receive interferon-based therapies (interferon-based treatment is not recommended for AP in the NCCN document). Patients with newly diagnosed BP may be enrolled in a clinical trial, undergo SCT, be treated with imatinib, or receive acute leukemia induction chemotherapy regimens (neither guideline recommends interferon). Imatinib is also a consideration in the relapsed or refractory disease settings when it has not previously been used.

When other treatment strategies have not been successful, chemotherapy with hydroxyurea or busulfan, transfusion support, or palliative care remain options for patients.

Implications for Future Research

Future directions of research on imatinib for CML fall into two main domains:

1. Clinical Sciences:
   • Efficacy of imatinib therapy alone or in combination with other agents.
   • Better predictors of patients most likely to respond or at risk of poor response.
   • Better understanding of the relative efficacy across segments of the population including different racial, ethnic and age groups.
   • Long-term longitudinal followup of imatinib in the various clinical settings.¹⁵⁵
   • Understanding of the ideal timing of SCT.
   • Meaning of surrogate markers such as molecular response at specific intervals after the initiation of therapy.
   • Impact of minimal residual disease when patients are in Complete CR.
   • Treatment algorithms subjected to objective evaluation.
   • Safe discontinuation of imatinib when there is a good clinical response.
   • Multiple drug regimens that include imatinib (Table 1d).
2. Basic Sciences:
   • Refined understanding of imatinib's mechanism of action (e.g., anti-angiogenic properties).
   • Molecular understanding of mechanisms of drug resistance for imatinib and other targeted therapies.
   • Better ability to predict individuals likely to be resistant to imatinib.
   • Development of new technologies so that knowledge of genetic profiles^156,157 and molecular predictors of resistance¹⁵⁸ can be translated into practical clinical tests.
   • Development of new targeted therapies that incorporate these molecular insights.

Standardization of terminology in CML is also important to advancing understanding of this disease. Blastic phase is a distinct period of the illness and some authors indicate that blast crisis is a sub-stage within blastic phase. This review highlighted the imprecision with which the terms blastic phase and blast crisis were used. A common language is needed to ensure that similar periods in the disease are compared across studies. Methods sections of manuscripts on CML should include a definition of how these terms are used.

Similarly, there are different definitions for the percentage of peripheral blood or bone marrow blasts that distinguish accelerated phase from blastic phase. Reviewers of this document suggested that 15 percent is the most commonly used cut-off. The NIH Web site, www.cancer.gov, cites 30 percent. Actual cut-off used across studies was variable. In order to ensure that assessment of efficacy by phase is accurate, it is critical that these definitions are standardized and that common terminology is used across studies. Methods sections should always include the definition.

Terminology for cytogenetic or clonal evolution is also imprecise. This is an important descriptor of the baseline participant population in a CML study and also considered by many authors as a predictor of disease response. Definitions of clonal evolution are rarely cited. Again, standardization of terminology and inclusion of definitions in methods sections is critical.

Return to Contents
Proceed to Next Section